Semiconductor devices are very small, typically from 5 mm square to 50 mm square, and typically comprise numerous sites for the bonding of electrical conductors to the semiconductor substrate. Each bond consists of a solder or gold ball deposit adhered to the substrate. It is necessary to test the strength of the attachment between the bond and the substrate, known as the bond strength, in order to be confident that the bonding method is adequate and that the bond strength is sufficient. Because of the very small size of the bonds, tools used to test the bond strength must be both very accurately positioned and able to measure very small forces and deflections.
A known test device, as described in WO2007/093799, has a test tool for engagement with a bond. The test tool is used to shear a bond off a semiconductor substrate and the force required to shear the bond is recorded. A force transducer is incorporated into the test tool in order to measure the force.
In order to ensure repeatability, it is essential for the tip of the test tool to engage the side of the bond at a predetermined height above the surface of the substrate. This distance is small but critical, since the bond is usually dome-shaped. A predetermined spacing from the surface eliminates both sliding friction from the test tool on the substrate, and ensures that the shear load is applied at a precise location in relation to the bond interface. Accordingly, in practice, the test tool is first moved into contact with the substrate surface and then withdrawn by a predetermined distance, typically 0.05 mm or less before the shear test is performed.
Several difficulties arise. Friction and stiction in the mechanism of the device itself may cause difficulties in sensing contact with the substrate surface. Imprecise surface sensing will inevitably affect the distance by which the test tool is withdrawn, and thus the height at which the bond is sheared. The distances involved are very small and so care needs to be taken to sense the exact moment of surface contact, without compression of the substrate. Care must also be taken to prevent uncontrolled movement of the test tool at the test height prior to or during the shear test. Such movement may seriously affect the accuracy of the test results and significant movement of the test tool at the test height may damage an adjacent bond or wire.
The twin objectives of both a low contact force when sensing the surface of the substrate and accurate control of the test height are difficult to resolve.
U.S. Pat. No. 6,078,387 discloses a device for sensing contact of a test head of a test tool with the substrate which is adapted to immediately stop downward drive of the test head when contact is sensed. The test tool is supported on the free end of a pair of cantilever arms which are secured at their opposite ends to a baseplate and deflects to allow some vertical movement of the test head with respect to the baseplate. To prevent vertical movement of the test head during the shear tests, test tool is spring biased by the cantilever arms against the baseplate. The test head can be moved away from the baseplate by an air-bearing to allow the test head to move vertically in a substantially frictionless manner for initial contact sensing. Thus, when the test head first touches the substrate surface, it is pushed back by the substrate surface on the cantilever arms. Movement of the test head relative to the baseplate or movement of the cantilever arm can be detected by an optical detector, and the air-bearing is then switched off to ensure that the test head is fixed relative to the baseplate by the spring bias of the cantilever arms against the baseplate. Once the test head is fixed relative to the baseplate, the baseplate is raised by a predetermined amount so as to leave a clearance between the lower end of the test tool and the substrate of the desired “step off distance”.
An alternative system is to have the cantilever arms bias the test tool away from the baseplate to allow for substantially friction-free movement of the test head relative to the base plate during initial positioning, but then to press the test tool into contact with the baseplate using a piston driven by compressed air to create a clamping force on the test tool against the baseplate during a test procedure.
Both of these systems are effective for accurately positioning a test head above a substrate whilst providing a relatively low touchdown force in initial positioning of the test tool above the substrate surface. However, they still suffer from some disadvantages.
A first disadvantage is the relatively high cost of these systems. The air-bearing is a relatively expensive component in the overall cost of a shear testing tool.
A second disadvantage is that during a test procedure much of the load has to be supported by the cantilever arms. This means that for tests at higher loads, larger, and consequently more massive, cantilever arms must be used. This in turn leads to a larger touchdown force when initially positioning the sensor, which can lead to damage to the substrate surface. This has also resulted in different cantilever arm assemblies being used for different load tests, further increasing cost.
A third disadvantage is that in both the air-bearing solution and the compressed air activated piston solution, the test head is moved laterally as the cantilever arms are urged or pressed against the baseplate, so that it can be positioned relative to the substrate but prior to the test being performed. The switching off of the air-bearing or the clamping of the test tool against the baseplate inevitably means some vertical movement will occur in addition to the lateral movement of the test head. This movement can reduce the accuracy of the resulting test, especially given the extremely small step off distance involved.
It is therefore an object of the invention to address the above-mentioned problems or at least to provide a useful alternative.